Non-verbal body pose communication classification on Non-verbal body pose communication dataset (test)
5.646Latency (ms)ST_GCN++
Evaluation results
| Method | Links | |
|---|---|---|
| ST_GCN++Params (M)=0.11, FP32 (MB)=0.48, FLOPs (M)=133.1, MACs (M)=66.5, Hardware=Jetson Orin Nano, Power profile=MAXN_SUPER, Software environment=PyTorch 2.5 + CUDA 12.6, Precision=FP32, Batch size=12026.06 | 5.646 | |
| AGCNParams (M)=0.07, FP32 (MB)=0.29, FLOPs (M)=122.8, MACs (M)=61.4, Hardware=Jetson Orin Nano, Power profile=MAXN_SUPER, Software environment=PyTorch 2.5 + CUDA 12.6, Precision=FP32, Batch size=12026.06 | 6.288 | |
| CTR_GCNParams (M)=0.06, FP32 (MB)=0.25, FLOPs (M)=114.7, MACs (M)=57.4, Hardware=Jetson Orin Nano, Power profile=MAXN_SUPER, Software environment=PyTorch 2.5 + CUDA 12.6, Precision=FP32, Batch size=12026.06 | 6.787 | |
| EfficientGCNParams (M)=0.03, FP32 (MB)=0.16, FLOPs (M)=30.4, MACs (M)=15.2, Hardware=Jetson Orin Nano, Power profile=MAXN_SUPER, Software environment=PyTorch 2.5 + CUDA 12.6, Precision=FP32, Batch size=12026.06 | 7.906 |